From Interview to Interface

An automated workflow that turns raw user interview data into finished Figma screens.

When the design process can no longer keep up

UX research and design processes are time-consuming. V-ZUG approached us with a vision to reduce this effort using AI, but without a concrete idea of what that might look like. At the time, automated AI persona generation from raw data and automated screen creation directly in Figma did not exist in this form. Deep research technology had only just been released, and there was no official AI write access for Figma yet. What was needed was not just a faster process, but a partner willing to experiment with the possibilities that were just emerging at the time.

V-ZUG

UI Design

UX Research

A process unfolds before your very eyes

Instead of spending weeks creating user personas and initial screen designs, we outlined a path for how these steps could be automated from raw interview data in the future. AI handles the core intermediate stages, while the final curation remains in human hands. When we presented this automation internally at V-ZUG, the reaction was clear: raw data was turned into personas and their user needs before the team's eyes, and from there into structured screens, without the many manual intermediate steps that the process previously required.

Raw data, organized.

The starting point was a series of qualitative interviews from a market survey. In the first step, we systematized how these raw data were handled, creating a foundation for everything that was to be automated subsequently.

AI personas at the touch of a button

Next, we connected two of our core tools: n8n and LeanScope.AI. We built the automation itself using n8n, while LeanScope.AI is used to create the personas and user scenarios.

Interview data is imported from several files into n8n and processed automatically. This means AI personas are generated automatically from the processed raw data—a novelty at the time.

Looking beyond the design process

In parallel with the technical implementation, we worked with V-ZUG to develop a roadmap for how AI can support the design and development process in the medium term, extending well beyond the hardware itself.

To achieve this, we took a close look at every process—from requirements gathering and processing to design workflows, engineering planning, and development—to identify areas where AI can effectively serve as a tool to augment human efforts.

From persona to screen

The automated workflow can be used today to take over multiple steps that were previously manual, from personas and scenarios to user needs and the first screen-design draft as a discussion basis directly in Figma. At the time, there was no official AI write access for Figma, which makes us especially proud that we achieved this with the tools available to us then. Today, we have many more possibilities available, which we would be happy to use in future collaborations.

A plan for the future

For the coming year – the near future – we defined with V-ZUG how AI can support the requirements process within a newly introduced requirements management tool, as well as the setup and maintenance of a design system.

A repeatable approach

This project is a prime example of how we approach automation at Centigrade: we identify a specific problem and solve it with precision, rather than using technology for its own sake. It is a principle that can be applied to other processes and projects as well.